There is enormous and growing interest in the consumption of up-to-the moment streams of newly-published content of various forms: news articles, posts on blogs, or bulletin boards, and multimedia data such as images, songs or movie clips. Users often consume such data on an as-generated basis, using mechanisms like atom and RSS to be notified when interesting content becomes available. Moreover, social media applications like flickr.com provide an opportunity for communities of users to build structure on top of base content using tags and annotations. In flickr.com for example, users may upload and share photos, and may place tags on their own or others' photos. Such an online image sharing service may allow a user to append a tag to any photo in the system resulting in the addition of over a million tags each week to the collection of photos accessible through the service. For any of these applications, understanding the evolution of such numerous tags presents a challenging task.
In order to explore the evolution of community focus for social media applications, there is a need for being able to browse through users, photos, tags, or more complex structures such as groups, themes and clusters. Past techniques for visualizing this information have been functional but inadequate. Schneiderman's Treemaps have been applied to evolving time-series data and provide a visualization of hierarchies. For example, SmartMoney's map of the web-based visualization shows multiple categories of time series data using a two-dimensional recursive partitioning of data points into boxes, and conveying volume and change in data using size and color. Unfortunately, this visualization focuses on a detailed breakdown of the data at each point in time without providing any framework for visualizing the evolution of the data overtime. Other approaches such as Moodstats show a static visualization of the evolution of mood over time, allowing detailed views into several dimensions of mood of an individual, and comparison to the snapshots of others. But the focus remains limited to providing a posthoc non-evolving view of an evolving dataset.
What is needed is a way to visualize the evolution of information built upon content more efficiently and that will apply at any timescale. Such a system and method should apply broadly to materializing and visualizing sequences of summarized data points along a time series for any type of content including audio, image, and video.